This is a repository copy of Polyamine functionalised ion exchange resins: Synthesis, characterisation and uranyl uptake.

White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/131854/

Version: Accepted Version

Article:

Amphlett, J.T.M., Ogden, M.D. orcid.org/0000-0002-1056-5799, Foster, R.I. et al. (3 more authors) (2018) Polyamine functionalised ion exchange resins: Synthesis, characterisation and uranyl uptake. Chemical Engineering Journal, 334. pp. 1361-1370. ISSN 1385-8947

https://doi.org/10.1016/j.cej.2017.11.040

eprints@whiterose.ac.ukhttps://eprints.whiterose.ac.uk/

Reuse

This article is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) licence. This licence only allows you to download this work and share it with others as long as you credit the authors, but you can’t change the article in any way or use it commercially. More information and the full terms of the licence here: https://creativecommons.org/licenses/

Takedown

If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

1

Polyamine Functionalised Ion Exchange Resins: Synthesis,

Characterisation and Uranyl Uptake

J. T. M. Amphletta, M. D. Ogdenb, R. I. Fostera,c, N. Synad, K. Soldenhoffd, *C. A. Sharrada

a School of Chemical Engineering and Analytical Science, The University of Manchester, Oxford Road,

Manchester, M13 9PL

b Separations and Nuclear Chemical Engineering Research (SNUCER), Department of Chemical and

Biological Engineering, The University of Sheffield, Mappin Street, Sheffield, S1 3JD

c Decommissioning Research Technology Division, Korea Atomic Energy Research Institute, Daejeon,

Republic of Korea

d ANSTO Minerals, Australian Nuclear Science and Technology Organisation, Locked Bag 2001,

Kirrawee D. C., NSW 2232, Australia

*Corresponding author: clint.a.sharrad@manchester.ac.uk

Abstract

A series of linear polyamine functionalised weak base anion exchange resins have been synthesised

using the Merrifield resin and characterised using infra-red spectroscopy, thermogravimetry,

elemental analysis and solid state 13C nuclear magnetic resonance spectroscopy. Uptake behaviour

towards uranium (as uranyl) from sulfuric acid media has been assessed as a function of pH and

sulfate concentration, with comparison to a commercially available weak base anion exchange resin,

Purolite S985. Synthetic polyamine resins were seen to outperform the commercial resin at

industrially relevant uranyl concentrations, with a trend of increased uptake being seen with

increasing polyamine chain length. Uranium loading isotherm studies have been performed and fit

with the Langmuir and Dubinin-Radushkevich isotherm models, with a maximum loading capacity

observed being 269.50 mg g-1 for the longest polyamine chain studied. Extended X-ray absorption

fine structure experiments have been used to determine uranium coordination environment on the

resin surface, showing a [UO2(SO4)3]4- species. This coordination knowledge was employed to

develop an extraction mechanism and derive an isotherm model based on the law of mass action.

Keywords; Uranium, polyamine, ion exchange, EXAFS, isotherm models

2

1. Introduction

As of the end of 2015, there are 441 nuclear reactors in operation around the world and 68

under construction [1]. The IAEA is projecting an increase in nuclear generating capacity worldwide

of up to 56% by 2030 and 135% by 2050 [1]. With this increasing global interest in nuclear power, its

continued sustainability is vital to meet the energy needs of the future. Current identified uranium

resources amount to 7,641,600 tU (tonnes of uranium metal), with reasonably assured resources

amounting to 4,386,400 tU, an increase on 2013 values of 0.1% for identified and a decrease of 4.4%

for reasonably assured resources [2]. Although these uranium resources are sufficient to meet the

increasing demand, low uranium prices due to poor market conditions may make economically

viable recovery difficult. This market stagnation is prompting interest in less conventional uranium

sources; however, these low prices have also caused a lack of investment in the development of new

uranium production technologies. This lack of development needs to be addressed, as current

processing flowsheets will not be suitable for these challenging conditions [2]. The adoption of new

techniques for the processing of less conventional uranium resources will result in the continued

sustainability of nuclear power through a reduction in fuel cost and therefore electricity price to the

public, as well as through a decreased environmental impact through extracting uranium as a co-

product and from tailings.

Current uranium milling processes employ either a sulfuric acid or carbonate based system

to solubilise the uranium, with the choice depending on the chemistry of the orebody [3に5]. Ideally,

the chosen lixiviant would only solubilise uranium as hexavalent uranyl (UO22+). However, due to the

nature of these dissolution processes, other deleterious species in the orebody (such as Fe, Al, Mg,

Mo, Mn, Ti, V, Zr) are usually dissolved into this leach liquor as well. The uranium needs to be

separated from these contaminants before it can be sold as a final concentrate product for

conversion, enrichment and fuel fabrication. The production of uranium from less conventional

resources is highly likely to exacerbate the problem of aqueous contaminants, as the

contaminant:uranium ratio will increase.

Currently, the favoured technologies for recovering uranium from the leach liquor are

solvent extraction (SX) and ion exchange (IX), though there is one example of nanofiltration being

used in a uranium process flowsheet [6]. For processes (i.e. in-situ or heap leach) where there will be

a relatively low concentration of uranium, IX is the preferred technique. This can be applied as a

standalone system, or as an upgrading system which can then be fed into further purification steps

(e.g. SX, precipitation). Additionally, there are numerous problems associated with SX which can be

overcome by using IX, such as; phase separation kinetics, solvent loss, generation of large volumes of

3

flammable and toxic organic waste, third phase formation and the need for an organic soluble

extracting ligand [7].

The removal of uranium from sulfate media by IX typically uses anion exchange resins. This is

due to the ability of uranyl ions to form anionic sulfate complexes (Eq. 1 - 3) [3,8,9] . Stability

constant values for complexation reactions are given in MHL stoichiometric format where notation K

indicates a stepwise equilibrium reaction while é Senotes the overall equilibrium reaction. Thus a

log10é value for the complexation of x metal ions, y protons and z ligands would be denoted as log10

éxyz.

戟頚態態袋 髪 鯨頚替態貸 蓋 戟頚態鯨頚替 詣剣訣怠待紅怠待怠 噺 に┻ひぱ Eq. 1 戟頚態態袋 髪 に 鯨頚替態貸 蓋 岷戟頚態岫鯨頚替岻態峅態貸 詣剣訣怠待紅怠待態 噺 に┻ぱぬ Eq. 2 戟頚態態袋 髪 ぬ 鯨頚替態貸 蓋 岷戟頚態岫鯨頚替岻戴峅替貸 詣剣訣怠待紅怠待戴 噺 ぬ┻ねの Eq. 3

There are two types of anion exchange resin, strong base and weak base. Strong base (SBA)

resins contain a quaternary ammonium functional group, with an associated anionic co-ion. These

are the type traditionally used in the mining industry [3,5,10]. Weak base (WBA) resins have also

been shown to effectively extract uranium from sulfuric acid based milling liquors [9,11,12], but their

use in uranium milling flowsheets has been limited. They can contain vinyl pyridine, polyamine and

polyamide functional groups [11]. Major advantages of these WBA resins are their selectivity for

uranium when compared to iron, a common and problematic contaminant element in uranium

processing, and a tolerance to dissolved chloride [5,9,11に14]. Though, due to their chemical nature,

these resins can only be implemented in liquors with pH 0 to pH 7 に 9, unlike SBA resins which can

tolerate much more alkaline conditions (pH 0 に 13) [9]. It is noted this does give SBA resins the

advantage of facile uranium elution in carbonate media. However, the advantages of SBA resins are

believed to outweigh their disadvantages, with tolerance to iron and chloride in solution becoming

increasingly important.

In this paper, the fundamental ion exchange characteristics of three polyamine

functionalised resins towards the uptake of UO22+ in sulfate media have been explored, with

comparison to a commercially available WBA resin, Purolite S985 (see Fig. 1 for functionality

structures) [15]. It is hypothesised that these homologous polyamine functional groups will show an

increase in uranium loading capacities with increasing chain length due to the increase in

protonatable nitrogen atoms. Work has also focussed on the determination of uranium speciation

on the surface of these resins, and, to the authorsげ knowledge, is the first paper to link a direct

measurement of this to an equilibrium loading isotherm model based on the mass action law.

4

Additionally, this work will begin to address the lack of data in available literature with regards to

uranium extraction from industrially relevant sulfate based leach liquors using WBA resins, with

literature database searches for publications concerning さion exchangeざ, さuraniumざ and さ┘W;ﾆ

H;ゲWざ returning less than 25 outputs. This will be the first in a series of papers regarding these

polyamine WBA resins, moving from initial screening and fundamental studies towards larger scale

pilot plant studies utilising consecutive loading and elution cycles.

Figure 1. Resin functional groups; Purolite S985 (a), Ps-EDA (b), Ps-DETA (c), Ps-PEHA(d) [15], R represents the resin

backbone structure.

2. Experimental

2.1. Reagents and Stock Solutions

Commercial IX resin Purolite S985 was supplied by Purolite. All reagents for resin synthesis

were supplied by Sigma Aldrich and used as received. Uranium solutions for uptake and extended X-

ray absorption fine structure (EXAFS) experiments were supplied by ANSTO Minerals and the

University of Sheffield, respectively. All resins were preconditioned by bottle rolling with 20 bed

volumes (BV) of 1 M H2SO4 for 24 hours and subsequently washed with 3 BV of deionised water (18

Mびぶ before use.

2.2. Synthesis

Merrifield resin (40 g, 5.5 mmol g-1 Cl) was left to swell in 1,4-dioxane (250 mL) for 24 hours.

EDA (16 g, 17.8 mL) and 1,4-dioxane (230 mL) was added to the swelled resin mixture and heated to

reflux under a nitrogen atmosphere with stirring for 48 hours. The resin was removed from the 1,4-

5

dioxane and then washed successively with ethanol (500 mL), triethylamine (10%) in

dichloromethane (300 mL), deionised water until washings reached pH 7, and ethanol (300 mL). The

resultant product was Ps-EDA. Synthesis of the other resins followed the same procedure, but with

different masses of the relevant amine reactant (Table 1).

Table 1. Masses of the amines EDA, DETA and PEHA used in resin synthesis.

Amine Mass / g

Ethylenediamine (EDA) 16.00

Diethylenetriamine (DETA) 27.70

Pentaethylenehexamine (PEHA) 61.86

2.3. Characterisation

Fourier transform infra-red (FT-IR) spectroscopy was performed using a Bruker ALPHA FTIR

spectrometer. Samples were prepared by grinding the resin and making a solid solution with KBr (1

wt %) followed by isostatic pressing into a pellet. FT-IR spectra of the pellets were then measured in

the transmission mode. Thermogravimetric analysis was carried out using a Mettler-Toledo

thermogravimetric analyser. A known mass of resin was added to a crucible and heated at a

constant rate of 20 °C min-1 from 25 °C to 800 °C. Solid state 13C NMR spectroscopy was performed

on ground, dry resin samples using a Bruker BioSpin 400 MHz NMR spectrometer. Elemental analysis

was carried out on a known mass of resin using a Thermo Scientific FLASH 2000 Series elemental

analyser.

2.4. Uptake Studies

For both pH and sulfate dependant uptake studies, solutions containing uranium (as UO22+)

at 1 g L-1 were used. All experiments were performed as batch extractions by contacting 2 mL of pre-

conditioned wet settled resin (WSR) with 50 mL of aqueous feed at room temperature for 24 hours

on an orbital shaker. The pH range studied was from -0.5 to 6. This was adjusted using sulfuric acid,

with sulfate concentration being dependent upon sulfuric acid concentration. The pH for all sample

solutions was monitored using a silver/silver chloride reference electrode calibrated from pH 1-10

using buffers. At high acid concentrations ([H+] > 0.1 M), [H+] was determined by titration with

standardised alkali solutions. Sulfate dependency studies were performed at a constant pH of 2, with

SO42- concentration (0 に 0.4 M) adjusted using Na2SO4. Extraction percentage of UO2

2+ was

determined by the difference between the uranyl solution concentrations pre- and post-uptake,

6

with ion concentrations determined by ICP-MS (Perkin Elmer Elan 9000) or by ICP-OES (Perkin-Elmer

Optima 5300DV).

2.5. Chloride Loading Capacity

An aqueous solution of HCl (1 M, 100 mL) was added to pre-conditioned resin (2 mLWSR) and

agitated for 24 hours. This process was repeated once and the resin was washed with deionised

┘;デWヴ ふヱΒ Mびが ン ┝ ヱヰ ﾏLぶく TｴW ヴWゲｷﾐ ┘;ゲ デｴWﾐ Iﾗﾐデ;IデWS ┘ｷデｴ ヱヰヰ ﾏL ﾗa HNO3 (2 M) and agitated for

24 hours. The mixture was allowed to settle and an aliquout (1 mL) of the solution phase was taken,

added to deionised water (4 mL), followed by the addition of potassium chromate (5 wt%, 0.1 mL).

This mixture was then titrated with a standardised aqueous solution of AgNO3 until a permanent red

Ag2CrO4 precipitate was seen.

2.6. Isotherm Studies

Loading isotherms were determined by contacting pre-conditioned resin (2 mLWSR) with

aqueous uranium solution (as UO22+; up to 10 g L-1 U, pH 2, 50 mL) by agitation on an orbital shaker

for 24 hours at room temperature. Aqueous uranium concentrations were determined pre- and

post-uptake by either ICP-OES or ICP-MS. Data were fitted to derived, Langmuir (Eq. 4) and Dubinin-

Radushkevich (Eq. 5) isotherm models, with fits being performed using the fitting function builder

and non-linear curve fitting in Origin Pro 2015 and standard errors calculated at 95% confidence

intervals:

圏勅 噺 長寵賑槌尿怠袋長寵賑 Eq. 4

圏勅 噺 圏陳結貸喋呑馴峭眺脹鎮津岾怠袋 迭頓賑峇嶌鉄 Eq. 5

where qe is [UO22+] on the resin at equilibrium, qm is the maximum [UO2

2+] loading capacity, Ce is

aqueous [UO22+] at equilibrium, b is the Langmuir constant, BDR is the Dubinin-Radushkevich

constant, R is the universal gas constant and T is temperature. These isotherm models are commonly

used to model uptake onto ion exchange resins, allowing extracted data to be compared with

previously published work. It is however important to note that due to the heterovalent nature of

the chemistry involved in ion exchange processes many of the underlying assumptions in the

theoretical derivations of these isotherm models are violated [9,16].

7

2.7. EXAFS Experiments

Uranium LIII-edge EXAFS spectra were recorded in transmission mode on beamline B18 at

the Diamond Light Source operating in a 10 min top-up mode for a ring current of 299.6 mA and an

energy of 3 GeV. The radiation was monochromated with a Si(111) double crystal, and harmonic

rejection was achieved through the use of two platinum-coated mirrors operating at an incidence

angle of 7.0 mrad. The monochromator was calibrated using the K-edge of an yttrium foil, taking the

first inflection point in the Y-edge as 17038 eV. Uptake was performed from a uranyl sulfate solution

(50 mL, 1 g L-1) in pH 2 H2SO4 using 2 mL of pre-conditioned ground resin. Sulfuric acid was used to

adjust solution pH where necessary. Resin was ground prior to uptake to produce a homogenous

sample and avoid packing related artifacts. Wet, dewatered resin (2 mL) was added to a cryo-tube

before being vacuum sealed in plastic. The samples were left this way during measurements.

3. Results and Discussion

3.1. Resin Characterisation

3.1.1 Elemental Analysis

Elemental analysis results for chloride and nitrogen content are presented in Table 2.

Theoretical nitrogen content has been calculated assuming that each polyamine chain only binds to

the resin at one point, therefore only displacing one chlorine atom. The yields produced using this

assumption do not agree with yields calculated from measured chlorine content in the resins. This

immediately infers the presence of crosslinking via polyamine chains and their attachment to

multiple benzyl groups. The amount of nitrogen in each resin increases as polyamine chain length

increases, however, the overall yields decrease with increasing polyamine chain length. The yield

calculated from the chlorine content does not change significantly between Ps-EDA and Ps-DETA,

indicating that both EDA and DETA molecules are capable of reacting with the same amount of

benzylchloride moieties. This yield is much lower for the Ps-PEHA resin. This is likely due to the

relatively large size of the PEHA molecule when compared with EDA and DETA restricting the

movement of it through the resin pores and reducing the amount of sites available for

functionalisation. Comparing the yields calculated from the nitrogen and chlorine contents allows for

the calculation of a value for the average number of bonds each polyamine chain has made to the

8

backbone structure of the resin. As expected, this value increases as polyamine chain length

increases. These values indicate a high propensity for crosslinking with these polyamine chains.

Table 2. Elemental analysis results for nitrogen and chlorine content in the Merrifield resin, Ps-EDA, Ps-DETA and Ps-

PEHA.

Nitrogen / % Chloride / %

Links per amine

Theoretical Measured Yield / % Theoretical Measured Yield / %

Merrifield 0.00 0.0 - 19.5 22.7 - -

Ps-EDA 17.90 9.82 54.9 0 2.18 90.4 1.65

Ps-DETA 26.85 11.54 43.0 0 2.41 89.4 2.08

Ps-PEHA 53.69 13.13 24.5 0 5.00 77.9 3.19

3.1.2 FT-IR Spectroscopy

Sections of the IR spectra (500 に 2000 cm-1) of the Merrifield resin, Ps-EDA, Ps-DETA and Ps-

PEHA are shown in Figure 2 (full spectra is available in Appendix 1, Fig.A.1, with full peak

assignments given in Tables A.1, 2, 3 and 4). The spectra of the synthetic resins are very similar,

though do differ from that of the Merrifield resin giving an initial qualitative insight that

functionalisation was successful. The synthetic method used to functionalise the Merrifield resin

with amine groups proceeds via a nucleophilic substitution reaction, with the amine replacing the

chloride. Therefore, stretching frequencies in the spectra associated with alkyl chloride groups

should not be observed upon functionalisation. The peak at 673 cm-1 in the spectrum of the

Merrifield resin has been assigned to a C-Cl stretch. This peak is not present in any of the

functionalised resin spectra. The Merrifield resin spectrum also shows a peak at 1265 cm-1, assigned

as a C-H wag in an RCH2Cl group which is not obviously present in the spectra of the synthesised

resins.

9

Figure 2. Section of the FT-IR spectra of the Merrifield resin, Ps-EDA, Ps-DETA and Ps-PEHA, with arrows denoting the C-Cl

stretch (673 cm-1) and the RCH2Cl stretch (1265 cm-1).

3.1.3 Thermogravimetric analysis

Plots of collected TGA data (Appendix 1, Fig.A.2) show a difference in the decomposition

profiles of each resin. The differences observed are due to different structures of the functionality

for each of the resins. There are three decomposition steps in each plot, the first of which (25 に 100

°C) relates to any residual water being driven off from the resins. Further steps pertain to the

breakdown of the resin itself. The extent of mass lost during the second step (230 - 380 °C) seems to

infer evidence of functionalisation. As the mass of the grafted functionality increases, so does the

percentage of mass lost from the resin, as would be expected. Mass losses are detailed in Table 3.

Chemical assignments of these mass losses are based on previously established studies where,

during the pyrolysis of anion exchange resins, nitrogen containing species were observed to be

liberated from the resin immediately above 200°C, and carbon containing species were seen to be

liberated at roughly 350°C [17,18].

10

Table 3. Mass loss during thermogravimetric analysis of the Ps-EDA, Ps-DETA and Ps-PEHA.

Temp. rangea / °C wt% lost Degradation Yield / %

Functionalisation / mmol g-1

Ps-EDA

29 に 135 4.6 Water

33.5 1.84 248 に 371 9.8 Polyamine/benzyl

371 に 655 68.4 Carbon matrix

Ps-DETA

25 に 109 3.8 Water

29.5 1.63 232 に 373 12.2 Polyamine/benzyl

373 に 686 70.5 Carbon matrix

Ps-PEHA

26 に 112 4.8 Water

44.1 2.42 231 に 383 27.0 Polyamine/benzyl

383 に 699 61.5 Carbon matrix a Temperature ranges were defined by determining the temperature where the gradient of the line deviates from that of

the intermediate slope between obvious mass losses.

An initial chloride loading on the starting Merrifield Resin of 5.5 mmol g-1 equates to 19.5 %

of the resin mass. This mass percentage should increase to 29.1, 41.3 and 61.4% for Ps-EDA, Ps-DETA

and Ps-PEHA respectively. However, such a large mass loss is not observed, indicating that there is

either incomplete conversion of benzylchloride groups to benzylamines, or that individual amine

chains have bonded to multiple benzyl moieties. Both of those scenarios are likely as not all the

chloride groups on the starting resin will be facing into the aqueous phase and able to react, as well

as there being multiple nitrogen attachment sites on each amine, PEHA in particular. Attempts have

been made to quantify the mass loss associated with the loss of polyamine/benzyl from TGA data (%

yield, functionality per gram of resin) which will allow for calculations of uranium:functionality ratios

from collected isotherm data. Calculated yields and resin functionality per gram of resin are shown

in Table 3. For these calculations it has been assumed that the entire mass loss assigned to this part

of the TGA profile for each resin is only associated with the polyamine chain for each resin and not

any other molecular constituents of these resins, allowing for a qualitative interpretation of the TGA

data. It was not possible to evaluate the species being cleaved from the resin, so it is likely that there

are large errors associated with the calculated yields, due to likely complex pyrolysis mechanisms

involved.

3.1.4 Solid State 13C NMR Spectroscopy

Solid state (SS) 13C NMR data of each of the synthesised polyamine resins and the base

Merrifield resin itself are provided in Table 4, with full spectra presented in Appendix 1 (A.3, 4, 5, 6).

Due to the nature of solid state NMR, and the high likelihood of crosslinking causing the formation of

more complex resin functionalities, complicated spectra were obtained for the polyamine resins

where full peak assignments, though attempted, must be considered tentative. In contrast, peaks in

the spectrum of the Merrifield resin are assigned with much more confidence but some

11

discrepancies with previous work have been noted. For example, the methylene carbon atom bound

to the chlorine in the Merrifield resin has been reported at both 46.2 and 40.1 ppm, with the

spectrum recorded in this work having peaks at both 39.98 and 45.76 ppm [19,20]. The spectrum

produced by Ansari et. al. assigns the methylene carbon to the peak at 40.1 ppm, but there is a peak

slightly upfield in the spectrum which has been left unassigned. These differences in peak

assignment between similar moieties illustrate the difficulty associated with this method for the

characterisation of solid polymeric resins. Further to this, a spectrum of polystyrene reported by

Joseph et. al. shows three peaks, at 146, 127 and 40 ppm, assigned to the quaternary aromatic

carbons, protonated aromatic carbons and aliphatic carbons, respectively [21]. This would suggest

that the peaks seen at 39.98 and 45.76 ppm in our Merrifield resin spectrum correspond to aliphatic

carbon atoms in the polystyrene-DVB matrix and the benzylchloride methylene carbon, respectively.

Our reported peaks at 127.59 and 145.29 ppm may therefore correspond to the quaternary and

protonated aromatic carbons, respectively. The smaller peak at 134.60 ppm may infer the presence

of sp2 alkene carbon atoms present from unreacted divinylbenzene.

Despite the difficulties in completely assigning the SS 13C NMR spectra of these resins, there

are obvious qualitative differences between the Merrifield and polyamine functionalised resin

spectra which must infer a difference in chemical structure, and therefore a successful

functionalisation reaction. The peaks in the aromatic region of the spectrum are relatively easy to

assign as they are at very similar chemical shifts to the initial resin. However, the peak assigned to

unreacted divinylbenzene groups in the Merrifield resin is only seen in the DETA functionalised resin,

at 137.85 ppm. This may be due to the peak being hidden in the EDA and PEHA spectra, or it could

be due to different batches of Merrifield resin being used in the initial synthesis. There is much more

variation seen at the downfield end of these spectra, with peaks being seen to broaden. These

correspond to the polymeric matrix and the carbon atoms associated with the amine groups. An

explanation for this broadening could be the effect of the protonation state of the nitrogen atoms

on the amine chain on the carbon chemical shift. SS 13C NMR shifts for EDA carbon atoms in the

aqueous phase have been reported at 43.9 and 37.6 ppm for the unprotonated and protonated

version, respectively [22]. As there are a large number of nitrogen atoms which can exchange

between protonated and deprotonated forms (particularly in the PEHA resin) it is likely that the

carbon atoms have unique chemical environments with subtly different chemical shifts. These slight

differences in chemical shift would lead to 13C NMR peak broadening and potential

overlapping/obscuring of other peaks.

12

Table 4. Solid state 13C NMR shifts and peak assignments for the Merrifield resin, Ps-EDA, Ps-DETA and Ps-PEHA.

Merrifield Ps-EDA Ps-DETA Ps-PEHA

Shift / ppm

Assignment Shift / ppm

Assignment Shift / ppm

Assignment Shift / ppm

Assignment

39.98 Aliphatic (resin

matrix) 40.45

Aliphatic (resin matrix)

41.06 Aliphatic (resin

matrix) 39.87

Aliphatic (resin matrix)

45.76 Methylene

(benzylchloride) 51.14

Aliphatic (benzylamino)

49.22 Aliphatic

(benzylamino) 57.87

Aliphatic (benzylamino)

127.59 Quaternary

aromatic 128.42

Quaternary aromatic

51.53 Aliphatic

(benzylamino) 129.00

Quaternary aromatic

134.60 sp2 alkene

(divinylbenzene) 146.48

Protonated aromatic

127.59 Quaternary

aromatic 146.34

Protonated aromatic

145.29 Protonated

aromatic - - 137.85

sp2 alkene (divinylbenzene)

- -

- - - - 143.92 Protonated

aromatic - -

Though all the evidence (FT-IR, TGA, NMR) points towards functionalisation having taken

place, it is not possible to know the exact molecular structure of the resin. It is highly likely that

individual polyamine chains (DETA and PEHA in particular) will have bound to the resin at multiple

sites, causing crosslinking. This could produce strong base sites, changing the character of the resin.

It is believed that the functionalisation on the resin is likely to be a mixture of multiple strong and

weak base functionalities. This may be advantageous, as this could convey the positive traits of SBA

resins and WBA resins, such as high loading capacity and selectivity. The fully quantitative

understanding of characterisation data from techniques applied to solid polymeric resins is

challenging, and this work shows how important it is to use multiple techniques which can

complement each other to produce a strong evidence base for proving the target product has been

successfully synthesised.

3.2. Uptake Studies

3.2.1 pH and Sulfate Dependant Uptake

The uptake of uranyl by each of the polyamine resins and Purolite S985 resin from aqueous

acidic sulfate media was explored using batchwise studies. All resins (2 mLWSR) were able to extract

over 85 % of uranyl from solutions (50 mL) with a starting uranium concentration of 1 g L-1 and at pH

1 に 6 ([SO42-] < 0.05 M) with 24 hours of contact time at room temperature (Fig. 3). Each of the linear

polyamine functionalised resins show a higher uranyl loading capacity relative to that observed from

the Purolite S985 resin. No dependency on the extent of uranyl uptake with respect to pH was

observed within this pH range for all studied resins. Increasing the [H+] beyond 0.1 M (pH 1) causes

the uranyl uptake to decrease dramatically for all studied resins (Fig. 3).

13

There is a clear trend in uranyl capacity between the linear polyamine resins, which follows

the increasing length of each polyamine functionality. This is suggestive of an increase in the number

of sites for uptake being present with increasing nitrogen content in the functionality, but the

branched functionality from the Purolite S985 does not fit with this simplistic interpretation so the

overall structure of the functionality does seem to influence uptake behaviour. At [H+]

concentrations above 0.3 M, the Purolite S985 resin has a higher affinity for uranyl than Ps-EDA and

Ps-DETA, although it still falls behind that of the Ps-PEHA resin. This is also seen in WBA resin DOWEX

M4195, where uptake suppression is seen to begin above 0.1 M H+ in H2SO4 media [23]. Studies into

amine functionalities grafted onto silica (aminopropylsilica, diethylenetriamine-propylsilica) showed

suppression of uranyl uptake at acidic pH, with less than 50% uptake at 0.05 M H+ (pH 3) [24].

Figure 3. Extent of uranyl uptake onto Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA with respect to [H+] in sulfuric acid

media ([U]ini (aq.) =1 g L-1; Vol.(aq) = 50 mL; Vol.WSR = 2 mL) . Black line fit is presented to assist guiding the eye and error

bars are ± 1 standard error.

The suppression of uranyl uptake by these resins is also seen as [SO42-] increases (Fig. 4),

which agrees with previously published work on WBA resins [23]. This effect can be explained using

chemical equilibria. The resins are preconditioned as the sulfate form, and it is this sulfate co-ion

which is exchanging with the anionic aqueous sulfate complex. As aqueous sulfate concentration

increases, the equilibrium position will tend towards the sulfate remaining on the resin, as ion

exchange becomes thermodynamically less favourable. There will also be competition between

anionic aqueous uranyl complexes and sulfate molecules for exchange sites on the resin. The general

14

trend in uranyl uptake with respect to the resins studied in this work, for the majority of the effluent

conditions studied where [H+] <0. 3 M and [SO42-] > 0.1 M is as follows:

Ps-PEHA > Ps-DETA > PsねEDA > S985.

Figure 4. Uranyl uptake onto Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA at pH 2 with increasing [SO42-] (as Na2SO4).

Black line fit is presented to assist guiding the eye and error bars are ± 1 standard error.

The strength of the interaction between the functionality and the extracted uranyl species

can be assessed by looking at the pH at which 50% extraction occurs (Table 5ぶく Tｴｷゲ さヮH50ざ I;ﾐ HW

extracted from the data by the polynomial fitting of the five closest data points to 50% uranyl

extraction (R2 > 0.9990). pH50 values indicate that the strength of the uptake interaction increases as

the polyamine chain length increases. However, this does not hold true for the Purolite S985 resin,

as it has the second highest pH50 but the lowest extraction percentage.

Table 5. pH50 and [H+]50 values for Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA.

Resin # N atoms pH50 [H+]50 / M p[SO42-]50 [SO4

2-]50 / M

Ps-EDA 2 0.064 0.862 0.680 0.209

Ps-DETA 3 -0.089 1.227 0.562 0.274

Ps-PEHA 6 -0.293 1.963 0.446 0.358

Purolite S985 6 -0.185 1.532 0.754 0.176

3.3. EXAFS Experiments

15

Multiple EXAFS data sets were collected for each resin. These were then combined and

normalised using the Athena software package [25]. These spectra were then processed and fits

were performed using the FEFF database via the software package Artemis [25,26]. There is little

visual difference between the spectra, alluding to the same uranium environment and uptake

mechanism for all resins (Fig. 5). Fits were attempted using crystallographic data from the inorganic

crystal structure database (ICSD) [27,28]. A chelation model was initially attempted with the

functional groups being multidentate with respect to the uranyl ion, however these did not give

statistically justifiable fits. When a uranyl species corresponding to an anion exchange mechanism

was fit, statistical parameters were acceptable. The fitted species is 6-coordinate in the equatorial

plane, with three bidentate sulfate groups ([UO2(SO4)3]4-), using a Hanning window with rmin = 1 Å

and rmax = 4 Å (Fig. 6). The U-S interatomic distances (Table 6) indicate a bidentate as opposed

monodentate sulfate group [29]. Fitting paths included one axial oxygen environment (n = 2), two

equatorial oxygen environments (n = 2, 4), two equatorial sulfate environments (n = 1, 2) and two

multiple scattering paths. The number of variables fit never exceeded two thirds of the independent

points available for each data set. This species fits with data collected for maximum resin capacity

measured by chloride uptake, as [UO22+]max/[Cl-]max Я ヴ ふT;HﾉW 7), indicating an anionic uranyl species

with -4 charge that interacts with the resins.

Figure 5. (A) The U LIII-edge k2-weighted EXAFS spectra and (B) the corresponding Fourier transforms (R-space) for uranyl

on Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA.

16

Figure 6. U LIII-edge EXAFS spectrum in R-space (above) and k-space (below) for the uptake of [UO2(SO4)3]4- onto synthetic

resin Ps-PEHA. Spectra for other resins can be found in Appendix 1 (A.7, 8, 9).

Table 6. U LIII-edge EXAFS data.

Scattering path na Interatomic Distances / Å ゝ2 b R-factorc

Purolite S985

U - Oaxial 2 1.79 0.00221

0.0183 U - Oequatorial 4 2.48 0.00488 U - Oequatorial 2 2.33 0.00358 U - S 2 3.11 0.00211 U - S 1 3.24 0.00176

Ps-EDA

U - Oaxial 2 1.79 0.00219

0.0204 U - Oequatorial 4 2.48 0.00444 U - Oequatorial 2 2.34 0.00356 U - S 2 3.11 0.00149 U - S 1 3.24 0.00108

Ps-DETA

U - Oaxial 2 1.79 0.00216

0.0213 U - Oequatorial 4 2.45 0.00444 U - Oequatorial 2 2.34 0.00337 U - S 2 3.11 0.00154 U - S 1 3.24 0.00127

Ps-PEHA

U - Oaxial 2 1.79 0.00230

0.0204 U - Oequatorial 4 2.48 0.00445 U - Oequatorial 2 2.34 0.00356 U - S 2 3.11 0.00147 U - S 1 3.24 0.00115

aNumber of occupants fixed. bDebye-Waller factor. cParameter describing the goodness of fit.

Table 7. The ratio of maximum uranyl capacity to maximum chloride capacity for Ps-EDA, Ps-DETA and Ps-PEHA,

determined by chloride titration.

Resin [UO22+]max / [Cl-]max

Ps-EDA 4.886

Ps-DETA 4.335

Ps-PEHA 4.068

17

3.4 Isotherm Modelling Studies

3.4.1 Langmuir and Dubinin-Raduchkevich

Isotherm data were fitted using both the Langmuir and Dubinin-Radushkevich isotherms;

with the goodness of fit parameter (R2) indicating a better fit was obtained with the Langmuir model

(Fig. 7, Tables 8 and 9). The trend in uptake capacity at low uranyl concentration is consistent with

pH and [SO42-] screening studies. However, at higher [UO2

2+] concentrations, the uranyl capacity of

Purolite S985 surpasses that of the synthetic polyamine resins. These high concentrations are

unrealistic when compared with those found in real mining circuits. The Dubinin-Radushkevich

isotherm model allows for the calculation of the mean free energy of sorption, where a value below

8 kJ mol-1 corresponds to a physisorption uptake mechanism, and one above corresponding to a

chemisorption uptake mechanism [23]. Although this isotherm model does not describe uptake

behaviour as well as the Langmuir, the mean free energy of sorption values obtained would suggest

a chemisorption mechanism, which agrees with an ion exchange system.

Table 8. Fitting parameters for Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA fit with the Langmuir isotherm model.

Ps-EDA Ps-DETA Ps-PEHA S985

qm / mg g-1 176.55 ± 2.51 195.47 ± 4.23 252.28 ± 4.46 283.22 ± 5.74

qm (x10-3) / mol g-1 0.742 ± 0.011 0.821 ± 0.018 1.060 ± 0.019 1.190 ± 0.024

b (x103) 1.097 ± 0.054 1.042 ± 0.080 1.397 ± 0.088 0.519 ± 0.036

R2 0.997 0.993 0.995 0.996

Table 9. Fitting parameters for Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA fit with the Dubinin-Radushkevich isotherm

model.

Ps-EDA Ps-DETA Ps-PEHA S985

qm / mg g-1 261.8 ± 14.87 295.12 ± 12.99 378.42 ± 18.69 480.76 ± 24.15

qm (x10-3) / mol g-1 1.100 ± 0.062 1.240 ± 0.055 1.590 ± 0.079 2.020 ± 0.101

BDR (x 10-9) 4.26 ± 0.28 4.40 ± 0.22 4.09 ± 0.23 5.81 ± 0.29

E / kJ mol-1 10.84 ± 0.35 10.66 ± 0.27 11.06 ± 0.32 9.28 ± 0.23

R2 0.971 0.983 0.978 0.986

18

Figure 7. Langmuir fits of isotherm data collected for Purolite S985, Ps-EDA, Ps-DETA and Ps-PEHA. The Langmuir model

fit for each resin is shown by a dashed line.

3.4.2 Derived Isotherm

Based on EXAFS experimental fits and chloride titration data, an isotherm model has been

derived. This isotherm is based on the equilibrium of two fully protonated [EDA(H)2]2+ functionalities

exchanging four bound [HSO4]- with one [UO2(SO4)3]4- species on the resin amine active site (Eq. 6).

This equilibrium allows for the derivation of an isotherm model representative of this IX system

based on the law of mass action (Eq. 7), where y=[UO2]resin, S=[UO2]resin,max, x=[UO2]aq and Kexげ=an

effective equilibrium constant (full derivation is available in Appendix 2). Experimental fits are

presented in Figure 8. Goodness of fit parameters (Table 10) suggest that this model fits the

collected data better than both the Langmuir and the Dubinin-Radushkevich isotherms.

に岷迎軽態茎態岫茎鯨頚替岻態峅 髪 岷戟頚態岫鯨頚替岻戴替貸峅 蓋 岷岫迎軽態茎態岻態岫戟頚態岫鯨頚替岻戴岻峅 髪 ね岷茎鯨頚替貸峅 Eq. 6

姿 噺 盤想皐蚕姉嫦 姉傘袋層匪罰謬盤貸想皐蚕姉嫦 姉傘袋層匪貸岫想皐蚕姉嫦 姉傘岻匝掻皐蚕姉嫦 姉 Eq. 7

19

Table 10. Fitting parameters for Ps-EDA, Ps-DETA, Ps-PEHA and Purolite S985 fit with the derived isotherm model.

Ps-EDA Ps-DETA Ps-PEHA S985

S / mg g-1 209.44 ± 6.75 233.24 ± 6.31 297.50 ± 5.67 365.33 ± 14.81

qm (x10-3) / mol g-1 1.760 ± 0.028 1.960 ± 0.027 2.500 ± 0.024 3.070 ± 0.062

Kexげ ふ┝ヱヰ5) 3.382 ± 0.239 2.837 ± 0.170 3.061 ± 0.132 0.755 ± 0.063

R2 0.997 0.998 0.999 0.997

Figure 8. Fitting of the derived isotherm to data collected for uranyl uptake onto Ps-EDA, Ps-DETA and Ps-PEHA and

Purolite S985. Speciation based isotherm model shown by dashed line fits.

Though the model fits with the data, it is likely to be a simplification of a complex

mechanism, potentially involving other uranyl species on the surface of the resin and in the bulk

solution. Even if in the bulk solution there are lower sulfated uranyl species (e.g. [UO2(H2O)x(SO4)y]

where y < 3) present, the expected high sulfate concentration within the immediate environment of

the sulfuric acid washed resins will most likely cause a change in any such uranyl species that

approaches these resins to the tris(sulfato) uranyl complex. As discussed previously, there may be

strong base sites produced from the crosslinking of amine chains and even unprotonated nitrogen

atoms which could interact directly with bound uranyl, potentially introducing a degree of chelation

to the mechanism. This could infer a degree of tolerance to high ionic strength (e.g. high Cl- in

20

seawater) which would be advantageous for these resins over traditional strong base resins. EXAFS is

a bulk technique giving an average of all species present, so the existence of a chelated species on

the resin cannot be ruled out completely. However, results do seem to suggest that the dominant

species on the surface of the resin is [UO2(SO4)3]4-, which gives confidence that this isotherm is

representative of uranyl speciation on these weak base anion exchange resins.

Observed maximum loading capacities for the Ps-EDA, Ps-DETA, Ps-PEHA and Purolite S985

are 172.73, 197.30, 256.21 and 269.50 mg g-1 respectively, which were achieved from a 10 g L-1

uranium solution. These values are predicted most effectively by the derived isotherm model, with

the Langmuir model, though fitting adequately, overestimates these values. WBA resin DOWEX

M4195 has been predicted to have a maximum loading capacity of 78 mg g-1 from Langmuir model

fits [23]. Anion exchange resin Amberlite CG-400 has shown predicted uptake capacities of 112.36

mg g-1 [30], with values for a ethylenediamine-tris(methylenephosphonic) acid resin being 64.26 mg

g-1 [31]. This shows the potential of these resins to outperform similar commercially available WBA

resins and ones with novel functionalities synthesised in academia (Table 11).

Table 11. Maximum uranyl capacities for a selection of ion exchange resins

Extractant qm (UO22+) / mg gr

-1 Aqueous Medium

Ps-EDA

172.73 pH 2 H2SO4

Ps-DETA

197.30 pH 2 H2SO4

Ps-PEHA

256.21 pH 2 H2SO4

Purolite S985

269.50 pH 2 H2SO4

Dowex M4195 [23]

78 pH 1.7 H2SO4

Amberlite CG-400 [30] 112.36 pH 3.5 HCl in the presence of PO43-

Polystyrene/Divinylbenzene functionalised with

ethylenediaminetris(methylenephosphonic) acid [32]

41.76 pH 3.4 nitric acid

Glycidylmethacrylate/Divinylbenzene functionalised with pentaethylenehexamine [33]

129.87 pH 4.5 nitric acid

Polystyrene/Divinylbenzene functionalised ┘ｷデｴ NがNげ-dimethyl-NがNげ-dibutylmalonamide [20]

18.78 3 M HNO3

Amberlite XAD-4 functionalised with succinic acid [34]

12.33 pH 4.5 Hexamine-HCl buffer

Polyethyleniminephenylphosphonamidic acid [35] 39.66 Nitric acid, concentration not given

The maximum uranium loading capacities predicted by the model fit can be used alongside

the estimated amount of functionality per gram of resin (Table 3) to calculate uranium:functionality

molar ratios and subsequently the uranium:nitrogen atomic ratios (Table 12). If it is assumed that all

21

the amine groups exist in the protonated form and no other ions apart from [UO2(SO4)3]4- are

present in the immediate binding sphere, the expected average uranium:functionality ratios would

be 0.50, 0.75 and 1.50 for Ps-EDA, Ps-DETA and Ps-PEHA, respectively, in order to maintain charge

balance. Despite the unlikely assumption that all the amine groups present on the resin are

protonated, these ratios indicate that there is more uranium present on the loaded Ps-EDA and Ps-

DETA resins than would be expected. We believe that the most likely reasoning for this is the use of

the TGA data is underestimating the total amount of functionality that is present on the resin. The

maximum chloride loading capacity for each resin falls within 4-5 times of the respective uranyl

capacity indicating the discrepancies in the uranyl:functionality ratios are unlikely to be due to errors

in determining the uranium capacities. The most likely explanation for the underestimation of the

functionality on the resins is the simplistic assumption that only the polyamine functionality is

cleaved during thermal degradation process is not valid. There is also the possibility that cross-linked

polyamine functionality groups do not degrade from the resin as readily as those attached at a single

site. This shows that the observed mass losses cannot be directly prescribed to the loss of a specific

functionality using TGA data alone and can only be used for a qualitative understanding under the

conditions studied. The combination of TGA with gas chromatography or mass spectrometry could

allow for an improved quantitative analysis. Repeating this process with collected elemental analysis

data (Table 2) produces uranium:functionality and uranium:nitrogen ratios which closely agree with

maximum loading capacities (Table 12).

Table 12. Calculated ratios of uranium:functionality (molar) and uranium:nitrogen (atomic) calculated by various means

using TGA and elemental analysis data for Ps-EDA, Ps-DETA and Ps-PEHA

Uranium:Functionality Uranium:Nitrogen

Calculated

theoretical

averagea

Experimental

calculation

(TGA)b

Experimental

calculation

(EA)c

Calculated

theoretical

averagea

Experimental

calculation

(TGA)b

Experimental

calculation

(EA)c

Ps-EDA 0.50 0.95 0.50 0.25 0.48 0.25

Ps-DETA 0.75 1.21 0.71 0.25 0.40 0.24

Ps-PEHA 1.50 1.03 1.60 0.25 0.17 0.27

a Calculated by assuming charge balance is maintained in the immediate binding sphere, all amine groups are protonated

and no other ions, apart from [UO2(SO3)]4-, are present in the immediate binding sphere.

b Calculated using the uranium loading estimated from derived model isotherm fits and the quantity of functionality

loading estimated from TGA data.

c Calculated using the uranium loading estimated from derived model isotherm fits and the quantity of functionality

loading calculated from elemental analysis data.

22

4. Conclusions The three synthetic resins (Ps-EDA, Ps-DETA and Ps-PEHA) have been successfully

synthesised as evidenced by the characterisation techniques employed. The Merrifield resin has

been shown to be an effective starting material from which to produce IX resins as the chloride

leaving group is able to participate readily in nucleophilic substitution reactions.

All synthetic resins were effective in removing uranyl from aqueous sulfate media, with

maximum uranyl loading capacities of 172.73, 197.30, 256.21 mg gwsr-1 observed for Ps-EDA, Ps-DETA

and Ps-PEHA, respectively. All synthetic resins outperformed the commercial resin Purolite S985 at

industrial process (i.e. in-situ or heap leach) uranyl concentrations (Table 14) but show lower qm

values at higher initial uranium concentrations. This suggests that Purolite S985 may be better suited

for uranium recovery from dynamic leach process liquors with higher uranyl concentrations [36]. It is

difficult to directly compare these results with those from other previously reported studies of

uranyl extraction by ion exchange resins due to the wide range of aqueous conditions used across

these investigations. Additionally, many of the uranyl maximum loading capacities reported are

predicted from isotherm models, not experimentally measured values.

Table 14. Constituents of a process water provided to the authors by ANSTO Minerals

Species Concentration / mg L-1

K+ 128

Na+ 107

Ca2+ 495

Mg2+ 5677

Mn2+ 2110

UO22+ 25

Al3+ 291

Fe3+ < 1

SO42- 12462

PO43- 6

SiO44- 31

EXAFS experiments suggested that the synthetic resins all follow a relatively simple ion

exchange mechanism, with [UO2(SO4)3]4- being extracted onto the resin. This mechanism fits the

collected isotherm data well, with calculated maximum uranium loading capacities from these fits

agreeing well with experimentally determined values.

These resins are promising candidates for further research towards their use in uranium

mining flowsheets. Future work with these resins will involve assessing their tolerance to common

contaminant species in uranium mining process waters, as well as the effects of Fe3+ and Cl- to

assess the potential application of these resins for the extraction of uranium from low quality ores

23

and waters, respectively. Determination of uranium extraction properties under continuous flow

column loading will be carried out in small scale laboratory experiments before scaling up to pilot

plant studies with multiple extraction/elution cycles. These studies will be presented in future

publications.

5. Acknowledgements

This work was funded by the UK Engineering and Physical Sciences Research Council (EPSRC

reference: EP/G037140/1). The authors would like to thank the Diamond Light Source for access to

beamline B18 (SP12643-1), especially Dr. Stephen Parry and Dr. Giannantonio Cibin for their

assistance with data acquisition, as well as Prof. Kath Morris for her help with sample preparation.

We thank ANSTO Minerals for hosting Mr. James Amphlett for a six month research visit, providing

expertise, aqueous uranium samples and access to analytical services. We also thank Dr. Barbara

Gore for her invaluable assistance with running SS 13C NMR experiments at the NMR service in the

School of Chemistry at The University of Manchester.

6. References

[1] IAEA, Energy, Electricity and Nuclear Power Estimates for the Period up to 2050, Vienna, 2016.

[2] Nuclear Energy Agency and International Atomic Energy Agency, Uranium 2016: Resources, Production and Demand, 2016. doi:https://www.oecd-nea.org/ndd/pubs/2004/5291-uranium-2003.pdf.

[3] C.R. Edwards, A.J. Oliver, Uranium processing: A review of current methods and technology, J. Miner. Met. Mater. Soc. 52 (2000) 12に20. doi:10.1007/s11837-000-0181-2.

[4] M.A. Ford, Uranium in South Africa, J. South African Inst. Min. Metall. 93 (1993) 37に58. doi:10.1111/j.1813-6982.1953.tb01589.x.

[5] A. Taylor, Short Course in Uranium Ore Processing, Short Course Uranium Ore Process. (2016).

[6] M. Peacock, S. McDougall, P. Boshoff, D. Butcher, M. Ford, S. Donegan, D. Bukunkwe, Nano-filtration technology for reagent recovery, in: ALTA Uranium-REE Sess., ALTA Metallurgical Services Publications, Perth, 2016.

[7] J. Veliscek-Carolan, Separation of actinides from spent nuclear fuel: A review, J. Hazard. Mater. 318 (2016) 266に281. doi:10.1016/j.jhazmat.2016.07.027.

[8] E.J. Zaganiaris, Ion Exchange Resins in Uranium Hydrometallurgy, Books on Demand France, 2009.

[9] E.M. Moon, M.D. Ogden, C.S. Griffith, A. Wilson, J.P. Mata, Impact of chloride on uranium(VI) speciation in acidic sulfate ion exchange systems: Towards seawater-tolerant mineral processing circuits, J. Ind. Eng. Chem. 51 (2017) 255に263. doi:10.1016/j.jiec.2017.03.009.

[10] International Atomic Energy Agency, Uranium Extraction Technology, Tech. Reports Ser. 359. (1993) 183.

[11] F.X. Mcgarvey, J. Ungar, The influence of resin functional group on the ion-exchange recovery of uranium, J. South. African Inst. Min. Metall. 81 (1981) 93に100.

24

[12] R. Kunin, R.L. Gustafson, E.G. Isacoff, H.F. Fillius, Ion Exchange Resins for Uranium Hydrometallurgy, Eng. Min. J. 170 (1969) 73に79.

[13] A. Rezkallah, J.F. Ferraro, P.M. Pellny, Removal of Uranium from Water, WO2014126699 A1, 2014.

[14] Boss Resources Limited, Boss Receives Positive Resin Test Results In Expansion Study, (2016). [15] A.N. Nikoloski, K.L. Ang, D. Li, Recovery of platinum, palladium and rhodium from acidic

chloride leach solution using ion exchange resins, Hydrometallurgy. 152 (2015) 20に32. doi:10.1016/j.hydromet.2014.12.006.

[16] M.A. Fernandez, G. Carta, Characterization of protein adsorption by composite silica-polyacrylamide gel anion exchangers. I. Equilibrium and mass transfer in agitated contactors, J. Chromatogr. A. 746 (1996) 169に183. doi:10.1016/0021-9673(96)00337-8.

[17] U. Chun, Pyrolysis and Oxidative Pyrolysis Experiments with Organic Ion Exchange Resins, Daejon, Republic of Korea, n.d.

[18] M. Matsuda, K. Funabashi, T. Nishi, H. Yusa, M. Kikuchi, Decomposition of Ion Exchange Resins by Pyrolysis, Nucl. Technol. 75 (1986) 187.

[19] J. Elguero, R.M. Claramunt, R. Garceran, S. Julia, M. Avila, J.M. del Mazo, 13C NMR Study of Polyphenyl-, Poly-N-azolyl- and Poly-N- benzazoyl-methanes, Magn. Reson. Chem. 25 (1987) 260に268.

[20] S. Aく Aﾐゲ;ヴｷが PくKく Mﾗｴ;ヮ;デヴ;が VくKく M;ﾐIｴ;ﾐS;が “┞ﾐデｴWゲｷゲ ﾗa NがNげ-dimethyl-NがNげ-dibutyl malonamide functionalized polymer and its sorption affinities towards U(VI) and Th(IV) ions, Talanta. 73 (2007) 878に885. doi:10.1016/j.talanta.2007.05.007.

[21] R. Joseph, W.T. Ford, S. Zhang, M.P. Tsyurupa, A. V Pastukhov, V.A. Davankov, Solid-State 13C-NMR Analysis of Hypercrosslinked Polystyrene, J. Polym. Sci. Part A Polym. Chem. 35 (1996) 695に701. doi:10.1002/(SICI)1099-0518(199703)35:4<695::AID-POLA12>3.0.CO;2-I.

[22] S.P. Dagnall, D.N. Hague, M.E. McAdam, 13C nuclear magnetic resonance study of the protonation sequence of some linear aliphatic polyamines, J. Chem. Soc. Perkin Trans. 2. (1984) 1111. doi:10.1039/p29840001111.

[23] M.D. Ogden, E.M. Moon, A. Wilson, S.E. Pepper, Application of chelating weak base resin Dowex M4195 to the recovery of uranium from mixed sulfate/chloride media, Chem. Eng. J. 317 (2017) 80に89. doi:10.1016/j.cej.2017.02.041.

[24] X. Sun, P.L. Zanonato, D. Bernardo, Z. Zhang, L. Rao, P. Di Bernardo, Sorption of Uranium and other Metal Ions on Amine-Functionalized Silica Materials Sorption of Uranium and other Metal Ions on Amine-Functionalized Silica Materials, Sep. Sci. Technol. 5018 (2015) 2769に2775. doi:10.1080/01496395.2015.1085403.

[25] B. Ravel, M. Newville, ATHENA , ARTEMIS , HEPHAESTUS票: data analysis for X-ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat. 12 (2005) 537に541. doi:10.1107/S0909049505012719.

[26] Mく NW┘┗ｷﾉﾉWが IFEFFIT票ぎ ｷﾐデWヴ;Iデｷ┗W XAF“ ;ﾐ;ﾉ┞ゲｷゲ ;ﾐS FEFF fitting, J. Synchrotron Radiat. 8 (2001) 322に324.

[27] MくBく Dﾗヴ;ﾐが AくJく Nﾗヴケ┌ｷゲデが Dく O げH;ヴWが E┝ヮﾉﾗヴ;デｷﾗﾐ ﾗa Cﾗﾏヮﾗゲｷデｷﾗﾐ “ヮ;IW ｷﾐ TWﾏヮﾉ;デWS Uranium Sulfates, Inorg. Chem. 42 (2003) 6989に6995. doi:10.1021/ic034540j.

[28] F.W. Lewis, L.M. Harwood, M.J. Hudson, M.G.B. Drew, M. Sypula, G. Modolo, D. Whittaker, C.A. Sharrad, V. Videva, V. Hubscher-Bruder, F. Arnaud-Neu, Complexation of lanthanides, actinides and transition metal cations with a 6-(1,2,4-triazin-3-yl)-ヲがヲ櫨ぎヶ櫨がヲ櫨櫨-terpyridine ligand: implications for actinide (III)/lanthanide(III) partitioning, Dalt. Trans. 41 (2012) 9209. doi:10.1039/c2dt30522d.

[29] C. Hennig, K. Schmeide, V. Brendler, H. Moll, S. Tsushima, A.C. Scheinost, The Structure of Uranyl Sulfate in Aqueous Solution - Monodentate Versus Bidentate Coordination, in: 13th Int. Proc. X-Ray Absorpt. Fine Struct., 2006: pp. 262に264.

[30] F. Semnani, Z. Asadi, M. Samadfam, H. Sepehrian, Uranium(VI) sorption behavior onto amberlite CG-400 anion exchange resin: Effects of pH, contact time, temperature and

25

presence of phosphate, Ann. Nucl. Energy. 48 (2012) 21に24. doi:10.1016/j.anucene.2012.05.010.

[31] V.K. Jain, A. Handa, S.S. Sait, P. Shrivastav, Y.K. Agarval, Preconcentration, separation and trace determination of lanthanum (III), cerium (III), thorium (IV) and uranium (VI) on polymer supported ovanillinsemicarbazone, Anal. Chim. Acta. 429 (2001) 237に246.

[32] A. Kadous, M.A. Didi, D. Villemin, A new sorbent for uranium extraction: Ethylenediamino tris(methylenephosphonic) acid grafted on polystyrene resin, J. Radioanal. Nucl. Chem. 284 (2010) 431に438. doi:10.1007/s10967-010-0495-7.

[33] S.A. Sadeek, E.M.M. Moussa, M.A. El-Sayed, M.M. Amine, M.O. Abd El-Magied, Uranium(VI) and Thorium(IV) Adsorption Studies on Chelating Resin Containing Pentaethylenehexamine as a Functional Group, J. Dispers. Sci. Technol. 35 (2014) 926に933. doi:10.1080/01932691.2013.809507.

[34] P. Metilda, K. Sanghamitra, J. Mary Gladis, G.R.K. Naidu, T. Prasada Rao, Amberlite XAD-4 functionalized with succinic acid for the solid phase extractive preconcentration and separation of uranium(VI), Talanta. 65 (2005) 192に200. doi:10.1016/j.talanta.2004.06.005.

[35] O. Abderrahim, M.A. Didi, D. Villemin, A new sorbent for uranium extraction: Polyethyleniminephenylphosphonamidic acid, J. Radioanal. Nucl. Chem. 279 (2009) 237に244. doi:10.1007/s10967-007-7270-z.

[36] D. Van Tonder, M. Kotze, Uranium recovery from acid leach liquors: IX or SX?, in: Alta 2007 Uranium Conf. Proc., 2007.